Machine Vision Toolbox for Python - data package

This package includes the images and data files used for examples with the Machine Vision Toolbox for Python (MVTB-P).

Rationale

The data files are provided as a separate package to work around disk space limitations on PyPI. Including these data with the MVTB code adds nearly 200MB to every release, which will blow the PyPI limit quite quickly.
Since the data doesn't change very much, it's mostly images models and a few data files, it makes sense for it to be a standalone package.

Package contents

Accessing data within the package

The Toolbox function mvtb_path_to_datafile(file) will return an absolute Path object that contains the path to file which is given relative to the root of the data package:

mvtb_path_to_datafile("images/monalisa.png")

which can also be used like os.path.join as

mvtb_path_to_datafile("images", "monalisa.png")

Image files are assumed to be in the images folder of the data package, and this will be searched by the image loading function

iread("myimage.png")       # read ./myimage.png
iread("monalisa.png")      # read from data package

or class method

Image.Read("myimage.png")       # read ./myimage.png
Image.Read("monalisa.png")      # read from data package

A matching local file takes precendence over a file in the data package.

Installing the package

You don't need to explicitly install this package, it happens automatically when you when you install MVTB-P

pip install machinevisiontoolbox-python

since it is a dependency.

Install really big image files

There are two very large zip files containing image sequences which are used in Sec. 14.8.3 Visual Odometry, each is 116M and exceeds the total PyPI quota. They are not included in the mvtbdata package, but you can download them into your local mvtbdata package by

python -m mvtbdata.mvtb_load_image_data

You only need to do this once.